Nils Everling |
| Nils graduated from KTH in 2017 with a M.Sc.Eng. in Computer Science, specialized in language technology. Film and data analysis are significant interests. Nils is currently employed as a quantitative analyst at AP4. |
Extending the explanatory power of factor pricing models using topic modeling - Computer Science Master thesis (2017) Link Abstract Factor models attribute stock returns to a linear combination of factors. A model with great explanatory power (R2) can be used to estimate the systematic risk of an investment. One of the most important factors is the industry which the company of the stock operates in. In commercial risk models this factor is often determined with a manually curated stock classification scheme such as GICS. We present Natural Language Industry Scheme (NLIS), an automatic and multivalued classification scheme based on topic modeling. The topic modeling is performed on transcripts of company earnings calls and identifies a number of topics analogous to industries. We use non-negative matrix factorization (NMF) on a term-document matrix of the transcripts to perform the topic modeling. When set to explain returns of the MSCI USA index we find that NLIS consistently outperforms GICS, often by several hundred basis points. We attribute this to NLIS' ability to assign a stock to multiple industries. We also suggest that the proportions of industry assignments for a given stock could correspond to expected future revenue sources rather than current revenue sources. This property could explain some of NLIS' success since it closely relates to theoretical stock pricing. |
 With topic modeling it is possible to capture multiple industry engagements for a given publicly listed company. The words listed in the chart distinguish topics, regardless of company.  A factor model with NLIS industry factors can explain a larger share of returns than a GICS-based model using the same number of factors (with equal or higher significance). 
A visual representation of the NLIS topic model. Reduction of 24 topics to 3 principal components. Colors denote GICS Industry Group - this displays whether NLIS assignments are consistent across stocks. Hover over a marker to see stock ticker and significant words. Link |
Cinetrii (2016-2017, 2020-) Link A search engine to trace artistic lineage in film based on inferences by critics. A software engineering hobby project making use of natural language processing. The best movie recommendation webapp we’ve come across - Huffington Post Depending on how you look at it, Cinetrii is a useful service, fun toy or dangerous time-suck. - Fandor Featured:  
 
 
|
HS2020 Energy monitor, project in DD1365 Introduction to Software Engineering (2014)Hammarby Sjöstad is running the sustainability project HS2020. I've been part of developing an application to visualize the energy consumption of the properties in the area. |
Final project in DD1339 Introduction to Computer Science (2013)With Timothy Bartram I developed a clone of Bomberman.
|
Game project (2014)I developed a 3D game engine in Java. Making the engine was far more fun than making the game, hence I never finished the game.
|